Delay-Constrained Video Transmission: Quality-Driven Resource Allocation and Scheduling

Abstract

Real-time video demands quality-of-service (QoS) guarantees such as delay\nbounds for end-user satisfaction. Furthermore, the tolerable delay varies\ndepending on the use case such as live streaming or two-way video conferencing.\nDue to the inherently stochastic nature of wireless fading channels,\ndeterministic delay bounds are difficult to guarantee. Instead, we propose\nproviding statistical delay guarantees using the concept of effective capacity.\nWe consider a multiuser setup whereby different users have (possibly different)\ndelay QoS constraints. We derive the resource allocation policy that maximizes\nthe sum video quality and applies to any quality metric with concave\nrate-quality mapping. We show that the optimal operating point per user is such\nthat the rate-distortion slope is the inverse of the supported video source\nrate per unit bandwidth, a key metric we refer to as the source spectral\nefficiency. We also solve the alternative problem of fairness-based resource\nallocation whereby the objective is to maximize the minimum video quality\nacross users. Finally, we derive user admission and scheduling policies that\nenable selecting a maximal user subset such that all selected users can meet\ntheir statistical delay requirement. Results show that video users with\ndifferentiated QoS requirements can achieve similar video quality with vastly\ndifferent resource requirements. Thus, QoS-aware scheduling and resource\nallocation enable supporting significantly more users under the same resource\nconstraints.\n